investigating saccharomyces cerevisiae alkene reductase

Investigating Saccharomyces cerevisiae Alkene Reductase OYE 3 by Substrate Profiling, X-Ray Crystallography and

Computational Methods

Journal: Catalysis Science & Technology

Manuscript ID CY-ART-03-2018-000440.R2

Article Type: Paper

Date Submitted by the Author: 27-Aug-2018

Complete List of Authors: Powell, III, Robert; University of Florida, Department of Chemistry Buteler, M. Pilar; University of Florida, Department of Chemistry Lenka, Sunidhi; University of Florida, Department of Chemistry

Crotti, Michele; Politecnico di Milano Dipartimento di Chimica Materiali e Ingegneria Chimica Giulio Natta Santangelo, Sara; Politecnico di Milano Dipartimento di Chimica Materiali e Ingegneria Chimica Giulio Natta Burg, Matthew; University of Florida, Department of Chemistry Bruner, Steven; University of Florida, Department of Chemistry Brenna, Elisabetta; Politecnico di Milano Dipartimento di Chimica Materiali e Ingegneria Chimica Giulio Natta Roitberg, Adrian; University of Florida, Department of Chemistry Stewart, Jon; University of Florida, Department of Chemistry

Catalysis Science & Technology

1

Investigating Saccharomyces cerevisiae Alkene Reductase OYE 3 by Substrate Profiling, X-Ray Crystallography and

Computational Methods

Robert W. Powell, III, M. Pilar Buteler†, Sunidhi Lenka†, Michele Crotti†,‡, Sara Santangelo†,‡, Matthew J. Burg, Steven Bruner, Elisabetta Brenna‡, Adrian E.

Roitberg and Jon D. Stewart*

Department of Chemistry, 126 Sisler Hall, University of Florida, Gainesville, FL 32611 USA

‡Dipartimento di Chimica, Materiali ed Ingegneria Chimica “Giulio Natta” Politecnico di Milano, Via Mancinelli 7, I-20131, Milano, Italy.

Phone & Fax 352.846.0743, E-mail [email protected]

*Author to whom correspondence should be addressed

†These authors contributed equally to this work

Page 1 of 43 Catalysis Science & Technology

2

Abstract

Saccharomyces cerevisiae OYE 3 shares 80% sequence identity with the well-studied

Saccharomyces pastorianus OYE 1; however, wild-type OYE 3 shows different

stereoselectivities toward some alkene substrates. Site-saturation mutagenesis of Trp 116 in

OYE 3 followed by substrate profiling showed that the mutations had relatively little effect,

opposite to that observed previously for OYE 1. The x-ray crystal structures of unliganded and

phenol-bound OYE 3 were solved to 1.8 and 1.9 Å resolution, respectively. Both structures were

nearly identical to that of OYE 1, with only a single amino acid difference in the active site

region (Ser 296 versus Phe 296, part of loop 6). Despite their essentially identical static x-ray

structures, molecular dynamics (MD) simulations revealed that loop 6 conformations differed

significantly in solution between OYE 3 and OYE 1. In OYE 3, loop 6 remained nearly as open

as observed in the crystal structure; by contrast, loop 6 closed over the active site of OYE 1 by

ca. 4 Å. Loop closure likely generates a greater number of active site protein contacts for

substrate bound to OYE 1 as compared to OYE 3. These differences provide an explanation for

the differing stereoselectivities of OYE 3 and OYE 1, despite their nearly identical x-ray crystal

structures.

Page 2 of 43Catalysis Science & Technology

3

Introduction

Since its discovery as the first cofactor-containing enzyme,1 the Old Yellow Enzyme from

Saccharomyces pastorianus2 (OYE 1) has been an important testbed for discovering general

principles of flavoprotein structure and mechanism.3-5 Biocatalytic interest in OYE 1 and its

relatives grew rapidly after Massey’s discovery that it reduces electron-deficient alkenes with

very high stereoselectivity.6 Over the past two decades, the substrate range of S. pastorianus

OYE 1 has been explored extensively; in parallel, many homologs have been cloned,

overexpressed and their synthetic abilities evaluated.7-9 In addition, a number of protein

engineering approaches have also been applied to OYE 1 and its homologs to improve substrate

acceptance, alter stereoselectivity, etc.10-14

OYE 1-mediated alkene reductions proceed with net trans-addition of both hydrogen atoms,

with hydride from reduced FMN adding to the substrate β-carbon lying above the cofactor and

protonation from the opposite face of the alkene mediated by a Tyr side-chain (Scheme 1).6

Substrate binding orientation thus dictates which face of the π system lies proximal to the flavin

and thereby determines the absolute stereochemistry of the reduced product.

Fox and Karplus solved the x-ray crystal structure of OYE 1 in 1994;15,16 since then, the

structures of several OYE 1 homologs have been determined.17-26 While all family members

share a common TIM barrel architecture and have similar active site structures, the observation

that some substrates bind in different orientations (and thus yield opposite stereoisomers)

underscores the importance of small structural differences (for an early example, see [Hall, 2008

#4544]).

The genomes of both bakers’ yeast (Saccharomyces cerevisiae) and brewers’ bottom yeast

(S. pastorianus) contain the OYE 1 gene.27 The former also includes two homologs (OYE 2 and


4

OYE 3).28 All share very high mutual sequence identity.29 To determine whether these minor

sequence differences impacted stereoselectivity, we examined all published examples of alkene

substrates that had been reduced by both OYE 1 and OYE 3. Unsurprisingly, the large majority

of alkenes were reduced with the same stereopreferences; however, five cases differed (Table 1).

In each, the reaction outcomes suggested that all five bound to OYE 1 in the “normal”

orientation (Scheme 1, left), but to OYE 3 in the “flipped” orientation (Scheme 1, right).

Because three of these examples involve substrates of particular interest to our synthetic

efforts,30 we solved the x-ray crystal structure of S. cerevisiae OYE 3. We also explored the

response of OYE 3 to mutations at Trp 116, a position that has proven highly influential in

dictating the stereoselectivity of OYE 1.31 Finally, we also carried out molecular dynamics

simulations of both OYE 1 and OYE 3 to uncover possible differences in active site structure and

dynamics that might help explain why reactions with OYE 3 sometimes differ from those of

OYE 1.

Experimental

General. Restriction endonucleases, Phusion Hot Start II High-Fidelity DNA Polymerase

and T4 DNA ligase were purchased from New England Biolabs and primers were obtained from

Integrated DNA Technologies. Crystallography screening kits (Classics Suite/ AmSO4 and

PEGRx HT) were purchased from Qiagen and Hampton Research, respectively. All other

reagents were obtained from commercial suppliers and used as received. Plasmids were purified

on small scales by Wizard® minicolumns (Promega Life Sciences) and on large scales using

CsCl density gradient ultracentrifugation. DNA sequencing was carried out by the University of


5

Florida ICBR using capillary fluorescence methods using standard protocols. Growth media

recipes have been described previously.[Sullivan, 2013 #5001]

Alkene substrates 5 – 19 were purchased from commercial suppliers or prepared in-house

during previous studies.33 Conversions and stereoselectivities were assessed by GC. Reactions

of substrates 3 and 4 were analyzed on a 0.25 mm × 30 m HP-5MS column (Agilent) coupled

with an MS detector to assess conversion, and on a 0.25 mm × 25 m Chirasil DEX CB column

(Chrompack) coupled with an FID to assess enantioselectivity. Reactions of substrates 5, 6, 10

and 13 - 19 were analyzed on a 0.25 mm × 30 m Beta Dex 225 column (Astec) coupled with an

FID. Reactions of alkenes 7 - 9 and 16 were analyzed on a 0.25 mm × 30 m DB-17 column

(Agilent J&W) coupled with an MS detector. Temperature programs were tailored to provide

baseline separations of each starting material and the corresponding product stereoisomers.

Determining the stereochemical courses of OYE 3 and OYE 1 reductions. Substrates 3 and 4

were prepared according to the literature.34,35 Racemic samples of 4-methylhexan-3-one36 and 4-

methylheptan-3-one,37 used as reference compounds, were prepared by hydrogenation of 3 and 4

on Pd/C in Et2O. The absolute configurations of 4-methylhexan-3-one and 4-methylheptan-3-

one, obtained by bioreduction, were established by comparison with optical rotation values

reported in the literature.38

(E)-4-Methylhex-4-en-3-one 3. 1H-NMR (CDCl3, 400 MHz):39 δ (ppm) 6.73 (1H, q, J = 6.9

Hz, CH=), 2.66 (2H, q, J = 7.3 Hz, COCH2CH3), 1.85 (3H, d, J = 6.9 Hz, CH3CH=), 1.78 (3H,

s, CH3C=), 1.09 (3H, t, J = 7.3 Hz, CH3CH2); 13C-NMR (CDCl3, 100.6 MHz):39 δ (ppm) 202.5,

138.2, 136.7, 30.4, 14.8, 11.2, 9.0; GC-MS (EI) m/z (%) = 112 (M+, 25), 83 (83), 55 (100).

(E)-4-Methylhept-4-en-3-one 4. 1H-NMR (CDCl3, 400 MHz):40 δ (ppm) 6.61 (1H, t, J = 7.2

Hz, CH=), 2.68 (2H, q, J = 7.3 Hz, COCH2CH3), 2.25 (2H, quintuplet, J = 7.3 Hz, CH2CH=),


6

1.78 (3H, s, CH3C=), 1.13 – 1.03 (6H, m, CH3CH2CO and CH3CH2C=); 13C-NMR (CDCl3, 100.6

MHz): δ (ppm) 202.6, 143.4, 136.6, 30.4, 22.4, 13.2, 11.3, 8.9; GC-MS (EI) m/z (%) = 126 (M+,

20), 97 (100), 69 (90).

4-Methylhexan-3-one. 1H-NMR (CDCl3, 400 MHz):36 δ (ppm) 2.51 – 2.40 (3H, m, COCH2

+ COCH), 1.75 – 1.60 (1H, m, CHHCH3), 1.45 – 1.33 (1H, m, CHHCH3), 1.06 (3H, d, J = 7.0

Hz, CH3CH), 1.05 (3H, t, J = 7.2 Hz, CH3CH2CO), 0.87 (3H, t, J = 7.6 Hz, CH3CH2CH);3 13C-

NMR (CDCl3, 100.6 MHz): δ (ppm) 215.4, 47.7, 34.4, 26.2, 16.1, 11.8, 7.9; GC-MS (EI) m/z

(%) = 114 (M+, 7), 86 (11), 57 (100).

4-Methylheptan-3-one. 1H-NMR (CDCl3, 400 MHz):37 δ (ppm) 2.58 – 2.50 (1H, m, COCH),

2.45 (2H, qd, J = 7.4 and 2.5 Hz, CH2CH3); 1.68 – 1.55 (1H, m, CHCHHCH2), 1.38 – 1.20 (3H,

m, CHCHHCH2 + CH2CH3), 1.06 (3H, d, J = 7.0 Hz, CH3CH), 1.04 (3H, t, J = 7.1 Hz,

CH3CH2CO), 0.90 (3H, t, J = 7.6 Hz, CH3CH2CH2); 13C-NMR (CDCl3, 100.6 MHz):37

δ (ppm)

215.5, 46.0, 35.4, 34.3, 20.6, 16.5, 14.2, 7.9; GC-MS (EI) m/z (%) = 128 (5), 86 (49), 71 (74), 57

(100).

Enone 4 was submitted to bioreduction with NADPD using both OYE 3 and OYE 1 in D2O.

The substrate (50 µmol) dissolved in i-PrOH-d8 (100 µL) was added to a KPi buffer solution (5.0

mL, 50 mM in D2O, pH 7.0) containing NADP+ (15 µmol), Thermoanaerobium brockii alcohol

dehydrogenase (4 U, 3 mg) and the required OYE (ca. 250 µg, dissolved in 500–700 µL H2O).

The mixture was incubated for 24 h in an orbital shaker (160 rpm, 30°C). The solution was

extracted with dichloromethane, centrifuging after each extraction, and the combined organic

solutions were dried over anhydrous Na2SO4. 2H spectra were recorded on a 400 MHz

spectrometer with proton broad band decoupling, in CHCl3 solutions, using CDCl3 as internal

reference for chemical shift scale.


7

Site-saturation mutagenesis of Trp 116 in OYE 3. Plasmid pRP4 (a derivative of pET 22b

with an OYE 3 coding region flanked by NdeI and XhoI sites) was the template used to make

OYE 3 Trp 116 site-saturation mutants. A pair of mutagenic primers containing a single codon

replacement for Trp 116 was used to make each of the 19 mutants. PCR was performed in a total

reaction volume of 50 µL, composed of 0.5 µL of template (18 ng / µL), 5 µL of both forward

and reverse primers (5 mM), 1 µL of dNTP mix (10 mM), 10 µL of 5 × HF Phusion® Hot start

buffer, 28 µL of sterile water, and 0.5 µL of Phusion® Hot Start II DNA Polymerase (2 U / µL).

PCR involved an initial denaturation step at 98 °C for 30 s, followed by 25 cycles of

denaturation at 98 °C for 10 s, annealing at 64 °C for 30 s, and an extension step at 72 °C for 3

min 30 s, followed by a final extension step at 72 °C for 7 min 30 s.

PCR products were purified by Wizard® Plus SV Gel PCR Clean up kits (Promega)

according to the manufacturer’s instructions. The first aliquot of DpnI (0.5 µL of a 20 U / µL

stock) was added; after incubating for 4 h at 37ºC, a second aliquot of DpnI was added and

incubation was continued overnight at 37ºC. Digested PCR products were purified as before,

then aliquots (4 µL) were used to transform 50 µL of E. coli ElectroTen-Blue cells (Agilent)

using 2.5 kV. Immediately following electroporation, 600 µL of SOC medium was added and

cells were incubated without shaking at 37 °C for 1 h, then plated onto LB plates supplemented

with 200 µg / mL ampicillin. Plates were grown overnight at 37ºC. Plasmids were isolated from

randomly-chosen colonies and sequenced to verify that the desired mutation was present, then

aliquots (4 µL) were used to transform 80 µL of E. coli BL21-gold(DE3) cells using

electroporation (2.5 kV). Immediately after electroporation, 600 µL of SOC medium was added

and cells were incubated without shaking at 37 °C for 45 min prior to plating onto LB plates

supplemented with 200 µg / mL ampicillin. Plates were incubated overnight at 37ºC. Plasmids


8

were isolated from each transformant and mutations were again verified by sequencing. The

library was arrayed in a 96 well microtiter plate and cells were grown overnight at 37ºC in 600

µL of LB medium supplemented with 200 µg / mL ampicillin. Aliquots (120 µL) of each

saturated culture were transferred to a new 96 well microtiter plate containing 30 µL of 80%

glycerol (which yielded a final glycerol concentration of 15%). This master plate was stored at -

80ºC.

Screening Trp 116 site-saturation mutagenesis library. E. coli BL21 (DE3) Gold cells

harboring plasmids encoding wild-type and Trp 116 site-saturation OYE 3 mutants were shaken

overnight at 250 rpm in a 96 well plate containing 600 µL of LB supplemented with 200 µg / mL

ampicillin at 37°C. Aliquots of each saturated culture (20 µL) were used to inoculate 2 mL wells

of a square bottom 96 well plate that contained 600 µL of ZYP-5052 auto-induction medium32

supplemented with 200 µg / mL ampicillin. Cultures were shaken overnight at 350 rpm and 37ºC

in a locally-designed and fabricated aeration case.41 Cells were harvested by centrifugation and

then resuspended in 600 µL of reaction mixture (50 mM KPi, 100 mM glucose, 15 mM alkene,

pH 7.0). Reaction mixtures were shaken overnight at room temperature at 250 rpm prior to

quenching with 500 µL of EtOAc. After thorough mixing, the organic phase was separated by

centrifugation and removed for GC analysis.

OYE 3 purification and crystallogenesis. OYE 3 protein was purified as described in the

Supporting Information. Fractions containing OYE 3 were pooled and concentrated by

ultrafiltration to a final concentration of 40 mg / mL (using A280 and an estimated value of ε280

(76,905 M-1cm-1).43

Initial crystallization conditions were identified by screening with the PEGRX HT (Hampton

Research) as well as the Classics Suite and AmSO4 Suite (Qiagen) kits. Sitting drop vapor


9

diffusion was used and wells contained 2 µL of crystallization solution and 2 µL of 20 mg / mL

OYE 3 (in 50 mM Tris-Cl, 50 mM NaCl, pH 7.5). The best crystals were obtained after 14 days

from well A8 of the PEG RX HT screening plate at room temperature and no further

optimization of these crystallization conditions was required (50 mM MES monohydrate, 22%

v/v PEG 400, 25 mM Tris, 25 mM NaCl, pH 8.0). Individual crystals were mounted and briefly

soaked briefly in cryoprotection buffer (100 mM MES monohydrate, 22% v/v PEG 400, 15%

(v/v) glycerol, pH 6.0) prior to flash cooling in liquid nitrogen for data collection.

Crystals were also grown from the well A10 conditions (PEG RX HT screening kit) using

sitting drop vapor diffusion (2 µL of 100 mM sodium citrate tribasic dihydrate, 30% (v/v)

polyethylene glycol monomethyl ether 550 crystallization solution and 2 µL of 20 mg / mL OYE

3 in 50 mM Tris-Cl, 50 mM NaCl, pH 7.5). Crystals grew after 10 days at room temperature.

They were mounted in appropriately-sized loops, then soaked briefly in cryoprotection buffer

(100 mM sodium citrate tribasic dihydrate, 30% (v/v) polyethylene glycol monomethyl ether

550, 2 15% (v/v) glycerol, pH 5.0) supplemented with with 2 mM p-hydroxybenzaldehyde.

Samples were flash cooled in liquid nitrogen prior to synchotron data collection.

Data collection and structure refinement. All data sets were collected on the 21-ID-G

beamline at the Advanced Photon Source, Argonne National Laboratory. Unsoaked and soaked

crystals diffracted to maximum usable resolutions of 1.8 and 1.9 Å, respectively. Reflection data

were processed using XDS.44 Phases were obtained by molecular replacement (PHENIX)45

using a modified S. pastorianus OYE 1 structure as the search model (PDB code 1OYB). All

ligands and water molecules were removed prior to molecular replacement. The initial model

was well defined throughout the scaffold of the protein except at the C-terminus, giving an Rfree

value of 0.37. Residues 398 and 399 (Lys and Asn, respectively) were poorly fit and these two


10

positions were therefore initially deleted from the model. The initially calculated 2Fo-Fc and Fo-

Fc maps showed electron density patterns readily identifiable as FMN. After an initial simulated

annealing refinement the Rfree dropped to 0.31. Residues 398 and 399 were added to the model

after the first round of refinement. Further refinement as well as continued cycles of model

building using the structure validation tools in COOT46 produced a final model with an Rfree

value of 0.22. Adding a chloride ligand in the active site accounted for most of the observed

active site electron density in the unsoaked crystals. For the p-hydroxybenzaldehyde-soaked

crystals, the ligand was readily identifiable by its electron density.

Molecular dynamics studies of OYE 3 and OYE 1. All simulations were performed using the

AMBER16 suite.47 The starting structure for OYE 1 was taken from the PDB (3TX9, 2.00 Å

resolution) and the OYE 3 starting structure was from this study. The Amber ff14SB force field

was used for the proteins and the parameters for the FMN cofactor were obtained from Bryce’s

group.48 Starting topologies were built using the tleap module of AmberTools 16.47 The systems

were solvated in a 10 Å buffer of TIP3P water49 using a truncated octahedral box under periodic

boundary conditions. The SHAKE algorithm50 was used to keep bonds involving H atoms at

their equilibrium length and an 8 Å cutoff was applied for nonbonded interactions. Newton’s

equations51 were integrated with a 4 fs time step, using hydrogen mass repartition52 and frames

were collected at 40 ps intervals. The collision frequency of the Langevin thermostat was 2 ps–1.

All simulations used different random seeds (ig = -1) to avoid synchronization artifacts.53

The initial structures were minimized for 5000 steps using steepest descent for the first 1000

cycles and switching to conjugate gradient, with 10 kcal mol–1Å–2 restraints applied to the

backbone. Heating was done using a linear temperature gradient from 10 to 300 K during 0.8 ns

at constant volume. Equilibration was performed at the final temperature during 8 ns using the


11

same backbone restraints under NPT conditions. Finally, we performed 100 ns of free dynamics

at a temperature of 300 K under NPT conditions. The simulations were done using

GPU(CUDA) version of the pmemd.cuda module.54 Five independent replicates of the

molecular dynamics runs were done to gauge the reproducibility of the results obtained.

Results and Discussion

Stereochemical courses of OYE 3- and OYE 1-mediated reductions

Our initial hypothesis for the stereochemical divergence between OYE 3 and OYE 1 was that

the two enzymes bound substrates in opposite orientations. Ketones with the general structure of

1 are reduced by both OYE 3 and OYE 1 with (S)-selectivity; this might arise from a “flipped”

substrate binding orientation in the active sites of both enzymes (Figure 1A). When the α-

substituent was lengthened by one methylene, e.g., ketones 2 and Table 1, entries 3 – 5,

reductions by OYE 3 and OYE 1 proceeded with divergent stereoselectivities (Table 1). This led

us to test 3 and 4, whose β-substituents (methyl and ethyl) are smaller than the aryl analogs

investigated previously. Would such compounds behave similarly to ketone 2, in spite of the

reduced steric hindrance of the β-group?

Trisubstituted alkenes 3 and 4 were reduced by wild-type OYE 3 and OYE 1 with (S)- and

(R)-stereoselectivities, respectively (Table 2), which imply “flipped” (OYE 3) and “normal”

substrate binding orientations (OYE 1). The alternative explanation – that OYE 3 and OYE 1

followed different stereochemical courses in H2 addition – was ruled out by reducing enone 4

with NADPD in D2O.55 In the 1H NMR spectrum of the racemic hydrogenation product from 4,

the two protons at C5 are characterised by different chemical shifts: one gives a multiplet in the


12

range 1.67-1.25 (shown in blue), and the other occurs as a multiplet in the range 1.38 -1. 20

(shown in red), which overlaps with the CH2 in position 6 (Figure 1B).

The 2H NMR spectra of the two deuterated samples show that both OYE 3 and OYE 1

promote the addition of two deuterium atoms in the same stereospecific way, presumably anti in

both cases (Figure 1C). The deuterium atoms are characterized by the same chemical shift, 2.45

and 1.21 ppm, corresponding to the multiplets at 2.44 and 1.38-1.20 ppm observed in the 1H

NMR spectrum of racemic reduction product of 4. No deuterium atom is present at the chemical

shift corresponding to the proton shown in blue (Figure 1B). These data demonstrate clearly that

the reversal of stereoselectivity between OYE 3 and OYE 1 does not arise because of a change

between net anti- and syn-addition of H2, but rather is a consequence of substrate binding

orientation in the active sites. Based on these results, we attempted to magnify differences in

enzyme-substrate interactions between OYE 3 and OYE 1 by mutagenesis of Trp 116, which has

a dramatic impact on binding orientation for some substrates in OYE 1.31,56 Would changes at

Trp 116 have the same impact on OYE 3? To answer this questions, all possible variants at

position 116 were created.57

Consequences of Trp 116 mutations in OYE 3 and OYE 1

We tested the Trp 116 mutant collections for both OYE 3 and OYE 1 (Table 2) against both

3 and 4. In general, replacements for Trp 116 had relatively little impact in OYE 3, the sole

exceptions being Phe for 3 and Leu, Met and Tyr for 4, which increased the amount of (R)-

product (although the overall stereoselectivity was near-racemic in all four cases). By contrast, it

was possible to identify mutants with either good (R)- and (S)-stereoselectivities in the Trp 116

replacement library of OYE 1. These results showed that OYE 1 is far more sensitive to active


13

site changes than OYE 3, suggesting that the active site of OYE 3 may have fewer protein-

substrate contacts that dictate binding orientation.

In previous work, we observed that the “flipped” binding mode is most commonly observed

with open chain trisubstituted alkenes.58 Only when the G group is large relative to the α-

substituent does the substrate adopt a “normal” binding orientation (Figure 1D). In the case of 3

and 4, however, the increased bulkiness of the ethyl (the G group) with respect to the methyl

group (the α-substituent) promotes the “normal” binding mode in OYE 1, but not in OYE 3, as it

was observed for ketone 2, thus confirming that the hindrance of the β-group has little influence.

Steric interactions between the G group and the residue at position 116 are much more important

in OYE 1; by contrast, OYE 3 strongly prefers a “flipped” substrate binding orientation for

trisubstituted alkenes, which cannot be overriden by changes to Trp 116.

To better define the importance of active site changes in OYE 3, the wild-type and all Trp

116 mutants were screened against 15 alkene substrates that we had investigated in previous

studies of OYE 1 using the same series of mutants (Figure 2).31,33,56

Wild-type OYE 3 was unable to reduce Baylis-Hillman adduct 5; by contrast, wild-type OYE

1 provides the reduction product from a “flipped” binding orientation with 60% ee (Table 3).56

For both OYE 3 and OYE 1, all Trp 116 variants reduced 5 from a “normal” substrate binding

orientation. A higher proportion of OYE 1 mutants retained catalytic activity toward 5 as

compared to OYE 3, although all stereoselectivities were consistently (S).

2-(Hydroxymethyl)cyclohexenone 6 provided an interesting contrast between OYE 3 and

OYE 1. While nearly all Trp 116 variants gave consistent (S)-stereoselectivities, Glu at position

116 of OYE 3 and OYE 1 yielded very different results (Table 3). In the sequence context of

OYE 3, the Trp 116 Glu mutant provided a nearly racemic mixture from Baylis-Hillman adduct


14

6 (2% ee favoring the (R)-enantiomer). The analogous OYE 1 mutant gave the (S)-product in

90% ee, consistent with all other Trp 116 variants in the sequence context of OYE 1.

Reductions of (S)- and (R)-carvone demonstrated even more clearly that altering Trp 116 in

OYE 3 and OYE 1 did not always yield the same results (Table 3). In the case of (S)-carvone 7,

the majority of substitutions for Trp 116 in OYE 1 prompted the alternative, “flipped” substrate

binding mode.31 This was not the case for OYE 3. In fact, none of the Trp 116 mutants

triggered the “flipped” substrate binding mode. This was a highly unexpected result. (R)-

Carvone 8 reductions also highlighted the subtle differences in active site behavior between OYE

1 and OYE 3. Two OYE 1 variants (Trp 116 Ala and Trp 116 Val) resulted in a predominantly

opposite stereochemical course for (R)-carvone reduction;31 however, the same two changes in

an OYE 3 background led to a nearly 1 : 1 mixture of diastereomers, showing that Trp 116

mutations exert less control over substrate binding orientation in OYE 3 as compared to OYE 1.

Neither wild-type OYE 3 nor any Trp 116 variant reduced (R)-pulegone 9 or

cyclopentenones 10 or 11 to significant extents. 2-Substituted cyclohexenone 12 was reduced by

wild-type OYE 3, but it was not examined as a substrate for the Trp 116 variants. As

expected,[Swiderska, 2006 #4350] sterically and electronically challenging 3-substituted

cyclohexenones 13 and 14 were recalcitrant toward reduction by wild-type or any of the OYE 3

Trp 116 variants. While many of the OYE 3 Trp 116 mutants gave reasonable conversion for

exocyclic enone 15 (with the Trp 116 Asn variant best), none yielded the opposite stereoisomer.

The series of 4,4-disubstituted cyclohexenones 16 – 19 probed the ability of OYE 3 to accept

bulky substituents in this portion of the substrate. Interestingly, while 16 and 17 were efficiently

reduced by wild-type OYE 3 and all Trp 116 variants, none accepted 18 and 19. While

reductions of 16 and 17 might show kinetic resolutions, this point was not investigated since the


15

starting enantiomers could not be separated by any of our available chiral GC columns;

moreover, the achiral nature of the products makes any kinetic resolution only marginally useful.

Crystallographic studies of OYE 3

To help understand these differences in stereochemical outcomes, we solved the x-ray crystal

structure of S. cerevisiae OYE 3. The protein was overexpressed in E. coli and purified by a

phenol affinity resin42 followed by gel filtration chromatography. After screening approximately

300 conditions, viable crystals were only observed when low molecular weight PEG was the

main precipitant at near neutral pH values (pH 6). The most successful conditions yielded

crystals that diffracted to a maximum usable resolution of 1.8 Å. The unit cell measured 61.2 ×

107.8 × 141. 1 Å and the crystals belonged to space group P212121 (Table 4). The asymmetric

unit contained two protein molecules and a solvent content of 53.5%. A Matthews coefficient of

2.65 Å3/Dal was calculated.

The structure of OYE 3 was solved by molecular replacement (MR) using OYE 1 as the

search model.15 Successive rounds of refinement and model improvement yielded an Rfree value

of 0.22. The overall TIM barrel structure of OYE 3 is nearly identical to that of OYE 1 (overall

rmsd = 0.26 Å using the 1OYB structure).

Once the protein scaffold was established, non-protein moieties were addressed. FMN was

easily fit into the structure using the observed electron density. The flavin environment of OYE

3 parallels that of OYE 1, with amino acids Thr 37, Gly 72, Gln 114 and Arg 243 hydrogen

positioned to form hydrogen bonds with the cofactor. The electron density in the vicinity of His

191 and Asn 194 was tentatively assigned to chloride as the predominant species. To help

understand how substrates might interact with OYE 3, we solved the structure after briefly

soaking with p-hydroxybenzaldehyde (Figure 3A). This had essentially no change on the OYE 3


16

structure (apart from displacement of the active site chloride ligand). The phenolate oxygen lies

between the side-chains of His 191 and Asn 194 and distances between the phenol and protein

residues for OYE 3 are very close to those observed previously for the analogous OYE 1

complex.15

All active site residues were identical between OYE 3 and OYE 1, except for position 296

(Ser in OYE 3 and Phe in OYE 1), located within loop 6 (residues 289 – 309). In native

(unsoaked) OYE 3 crystals, the electron density for all residues in loop 6 is very well-defined

(Figure 3B). This is also the case for OYE 3 crystals soaked with p-hydroxybenzaldehyde

(Figure 3C), although three solvent-exposed Glu side-chains showed somewhat weak electron

density beyond Cβ.

Structural comparison of OYE 3 and OYE 1

As expected from their high sequence identity, the overall structures of OYE 3 and OYE 1

are highly similar. Even loop 6, which contains the only active site change (Ser 296 in OYE 3

and Phe 296 in OYE 1), was found in the essentially the same conformation for OYE 3 (in both

ligand-soaked and unsoaked crystals) as well as in ligand-soaked crystals of OYE 1 (Figure 4).

The only significant difference between OYE 3 and OYE 1 in crystallo is the rotameric

preference of the side-chains at position 296. In OYE 1, Phe 296 extends into the active site; in

OYE 3, its side-chain is directed outward toward the solvent (Figure 4).

Computational studies of OYE 3 and OYE 1

Since substrate binding orientation was the key to dictating stereoselectivity, we used the

relevant crystal structures to model Michaelis complexes for all five substrates in Table 1 into

the active sites of OYE 3 and OYE 1. Each was modeled separately in the “normal” and

“flipped” orientations. This approach had been successful in understanding the stereopreferences


17

in OYE 1 and other homologs.41 For both enzymes, the “normal” substrate binding mode was

easily modeled by overlaying the dashed lines with the p-hydroxybenzaldehyde bound to the

active sites (Table 1). In the models of “flipped” substrate binding, the phenolic oxygen of p-

hydroxybenzaldehyde and the meta-ring carbon marked the approximate locations of the

carbonyl oxygen and β-carbon of the substrates, respectively (Scheme 1). In all cases, the

substrate positions were consistent with angle and distance values observed for efficient hydride

transfer from reduced FMN to Cβ,59 supporting the notion that these modeled structures might

mirror catalytically productive complexes. Key protein-substrate distances were measured to

identify possible steric interactions that would influence the choice “normal” versus “flipped”

substrate binding.

Unfortunately, the modeled Michaelis complexes utterly failed to uncover any explanations

for the differences in stereopreferences between OYE 3 and OYE 1. In fact, our models based

on the x-ray structures predicted (incorrectly) that both enzymes would give entirely consistent

outcomes, contradicting the experimental observations.

Because the active sites of OYE 3 and OYE 1 differ at position 296, and this residue lies on

mobile loop 6, we considered the possibility that these proteins might possess different

conformational and / or dynamic properties in solution that were not apparent in the x-ray

structures. Each crystal structure was used as the starting point for MD simulations in the

absence of substrate. Over the course of the simulations, variations in loop 6 conformations

were observed; however, the average structures were well-defined (Figures 5A and 5B). While

both proteins showed approximately the same level of conformational mobility during the

simulations, the average structure of loop 6 was more closed in OYE 1 as compared to OYE 3.

This can be seen in the distance distributions between the Cα of residue 296 and the site of


18

hydride addition (N5 of the FMN) (Figure 5C). In the starting x-ray structures of OYE 3 and

OYE 1, these atoms are 10.1 and 9.9 Å apart, respectively. The MD simulation shows that in

solution, loop 6 of OYE 1 closes significantly since the Cα,296 – N5,FMN distance decreased to

approximately 7 Å. By contrast, loop 6 in OYE 3 failed to close over the active site – in fact, the

Cα,296 – N5,FMN distance actually increased slightly to ca. 11 Å (Figure 5D). The more open

conformation of loop 6 over the active site of OYE 3 likely means that it has negligible contact

with bound substrates, which would lessen its impact on stereoselectivity. Based on these

results, we anticipate that the significantly altered loop 6 conformation in OYE 1, which is

largely absent in OYE 3, is the major reason for the stereochemical differences that have been

observed between these two enzymes (Table 1).

In the crystal lattices, loop 6 in both OYE 3 and OYE 1 faces a solvent channel, and there are

no close contacts with symmetry-related protein molecules. Thus, the nearly identical loop 6

conformations observed in crystallo for these two proteins are not likely to be due to crystal

packing forces. As noted above, the electron density for loop 6 is very well-defined in both the

OYE 3 and OYE 1 structures and B-factors in this region are nearly the same as for other surface

loops. Only by carrying out MD simulations in solution did the conformational differences

become apparent.

That loop 6 is more closed over the active site in OYE 1 versus OYE 3 is also consistent with

the stereochemical behavior of the two enzymes. In general, OYE 1 provides greater chiral

discrimination, and changes elsewhere in the active site, e.g., position 116, had large impacts on

stereoselectivity. By contrast, the lessened ability of loop 6 to close over the OYE 3 active site

yields a more open cavity with multiple opportunities for non-productive enzyme-substrate


19

complexes to form. Moreover, this means that contacts with other amino acid side-chains will

have lessened ability to influence substrate binding, as observed in this study.

Finally, the results from circular permutation studies of OYE 1 by the Lutz group are also

consistent with our observations here.13,60 The largest increases in catalytic efficiency (≥ 10-

fold) occurred when the new termini dissected loop 6, which created an effectively “permanently

open” loop 6. Interestingly, this group also found that mutating Trp 116 in the most catalytically

efficient circular permutants had little impact on the stereoselectivity of (S)-carvone reduction,

reminiscent of our observations on OYE 3. In both cases, loop 6 occupies a more open

conformation than that found in wild-type OYE 1.

Conclusions

The conformation of loop 6 in S. pastorianus OYE 1 likely has a different structure in

solution than that observed in its crystal structure. In solution, loop 6 of OYE 1 closes down

over the active site, allowing it to interact closely with bound substrates and exert significant

control over their binding orientations. By contrast, loop 6 in S. cerevisiae OYE 3 largely retains

its crystallographically observed conformation in solution. Because the loop does not close over

the active site of OYE 3, there are fewer protein – substrate contacts and substrate binding is less

impacted by residues on this loop and elsewhere in the active site, e.g., Trp 116. The divergent

conformational preferences of loop 6 in OYE 1 and OYE 3 in solution may explain their

differing stereochemical preferences toward some substrates, despite the two proteins sharing

80% sequence identity and nearly the same crystal structures.

Whether the loop 6 conformations of OYE 3 and / or OYE 1 can be further impacted by

additional mutations must await future experimental and computational studies that are currently


20

underway. While in principle, mutating Ser 296 to Phe in OYE 3 might be expected to yield a

variant with properties identical to OYE 1, a number of other residues differ between OYE 3 and

OYE 1, and a single amino acid change may not be sufficient to effectly turn one protein into

another. Only by examining a complete set of mutations and molecular dynamics trajectories

will conclusions emerge. We also note that such altered structures might have very useful

impacts on stereoselectivity, particularly when combined with additional active site changes.

What has been clearly established by this work, however, is that static x-ray structures – while

extremely useful – are not always sufficient to understand OYE-mediated alkene reductions.

Conflicts of Interest

There are no conflicts of interest to declare.

Acknowledgements

We are very grateful for financial support from the National Science Foundation (CHE-

1111791 and CHE-1705918). We would also like to thank Blue Waters (National Center for

Supercomputing Applications) and the High Performance Computing Center (University of

Florida) for providing computational resources.

Electronic Supplementary Information

Complete NMR spectra for ketones 3 and 4 along with their reduced products.


21

References and Notes

(1) O. Warburg and W. Christian Naturwissenschaften 1932, 20, 980-981.

(2) Formerly known as Saccharomyces carlsbergensis.

(3) A. S. Abramovitz and V. Massey J. Biol. Chem. 1976, 251, 5327-5336.

(4) R. G. Matthews and V. Massey J. Biol. Chem. 1969, 244, 1779-1786.

(5) R. G. Matthews, V. Massey and C. C. Sweeley J. Biol. Chem. 1975, 250, 9294-9298.

(6) A. D. N. Vaz, S. Chakraborty and V. Massey Biochemistry 1995, 34, 4246-4256.

(7) M. Hall and A. S. Bommarius Chem. Rev. 2011, 111, 4088-4110.

(8) K. Durchschein, M. Hall and K. Faber Green Chem. 2013, 15, 1764-1772.

(9) H. S. Toogood and N. S. Scrutton Curr. Opin. Chem. Biol. 2014, 19, 107-115.

(10) D. J. Bougioukou, S. Kille, A. Taglieber and M. T. Reetz Adv. Synth. Catal. 2009, 351,

3287-3305.

(11) M. Hulley, H. S. Toogood, A. Fryszkowska, D. Mansell, G. M. Stephens, J. M. Gardiner

and N. S. Scrutton ChemBioChem 2010, 11, 2433-2447.

(12) H. Toogood, A. Fryszkowska, M. Hulley, M. Sakuma, D. Mansell, G. M. Stephens, J. M.

Gardiner and N. S. Scrutton ChemBioChem 2011, 12, 738-749.

(13) A. B. Daugherty, S. Govindarajan and S. Lutz J. Am. Chem. Soc. 2013, 135, 14425-14432.

(14) E. D. Amato and J. D. Stewart Biotechnol. Adv. 2015, 33, 624-631.

(15) K. M. Fox and P. A. Karplus Structure 1994, 2, 1089-1105.

(16) P. A. Karplus, K. M. Fox and V. Massey FASEB J. 1995, 9, 1518-1526.

(17) S. Horita, M. Kataoka, N. Kitamura, T. Nakagawa, M. T., J. Ohtsuka, K. Nagata, S.

Shimizu and M. Tanokura ChemBioChem 2015, 16, 440-445.


22

(18) T. Barna, H. L. Messiha, C. Petosa, N. C. Bruce, N. S. Scrutton and P. C. F. Moody J. Biol.

Chem. 2002, 277, 30976-30983.

(19) S. Reich, H. W. Hoeffken, B. Rosche, B. M. Nestl and B. Hauer ChemBioChem 2012, 13,

2400-2407.

(20) Y. A. Pompeu, B. Sullivan, A. Z. Walton and J. D. Stewart Adv. Synth. Catal. 2012, 354,

1949-1960.

(21) A. Fryszkowska, H. Toogood, M. Sakuma, G. M. Stephens, J. M. Gardiner and N. S.

Scrutton Catal. Sci. Technol. 2011, 1, 948-957.

(22) D. J. Opperman, B. T. Sewell, D. Litthauer, M. N. Isupov, J. A. Littlechild and E. van

Heerden Biochem. Biophys. Res. Commun. 2010, 393, 426-431.

(23) B. V. Adalbjörnsson, H. Toogood, A. Fryszkowska, C. R. Pudney, T. A. Jowitt, D.

Leys and N. S. Scrutton ChemBioChem 2010, 11, 197-207.

(24) B. G. Fox, T. E. Malone, K. A. Johnson, S. E. Madson, D. Aceti, C. A. Bingman, P. G.

Blommel, B. Buchan, B. Burns, J. Cao, C. Cornilescu, J. Doreleijers, J. Ellefson, R.

Frederick, H. Geetha, D. Hruby, W. B. Jeon, T. Kimball, J. Kunert, J. L. Markley, C.

Newman, A. Olson, F. C. Peterson, G. N. Phillips, Jr., J. Primm, B. Ramirez, N. S.

Rosenberg, M. Runnels, K. Seder, J. Shaw, D. W. Smith, H. Sreenath, J. Song, M. R.

Sussman, S. Thao, D. Troestler, E. Tyler, R. Tyler, E. Ulrich, D. Vinarov, F. Vojtik, B. F.

Volkman, G. Wesenberg, R. L. Wrobel, J. Zhang, Q. Zhao and Z. Zolnai Proteins: Struct.

Func. Bioinformatics 2005, 61, 206-208.

(25) T. E. Malone, S. E. Madson, R. L. Wrobel, W. B. Jeon, N. S. Rosenberg, K. A. Johnson, C.

A. Bingman, D. W. Smith, G. N. Phillips, Jr., J. L. Markley and B. G. Fox Proteins:

Struct. Func. Bioinformatics 2005, 58, 243-245.


23

(26) K. Kitzing, T. B. Fitzpatrick, C. Wilken, J. Sawa, G. P. Bourenkov, P. Macheroux and T.

Clausen J. Biol. Chem. 2005, 280, 27904-27913.

(27) K. Saito, D. J. Thiele, M. Davio, O. Lockridge and V. Massey J. Biol. Chem. 1991, 266,

20720-20724.

(28) K. Stott, K. Saito, D. J. Thiele and V. Massey J. Biol. Chem. 1993, 268, 6097-6106.

(29) OYE 1 and OYE 2 show 92.0% overall identity while OYE 1 and OYE 3 are 80.5%

identical.

(30) E. Brenna, S. L. Cosi, E. E. Ferrandi, F. G. Gatti, D. Monti, F. Parmeggiani and A.

Sacchetti Org. Biomol. Chem. 2013, 11, 2988-2996.

(31) Y. A. Pompeu, B. Sullivan and J. D. Stewart ACS Catal. 2013, 3, 2376–2390.

(32) F. W. Studier Protein Express. Purif. 2005, 41, 207-234.

(33) A. Patterson-Orazem, B. Sullivan and J. D. Stewart Bioorg. Med. Chem. 2014, 22, 5628-

5632.

(34) D. Kalita, M. Morisue and Y. Kobuke New J. Chem. 2006, 30, 77-92.

(35) D. D. Faulk and A. Fry J. Org. Chem. 1970, 35, 364-369.

(36) H. C. Brown, R. K. Bakshi and B. Singaram J. Am. Chem. Soc. 1988, 110, 1529-1534.

(37) L. Wang, Z. Xu and T. Ye Org. Lett. 2011, 13, 2506-2509.

(38) For (R)-4-methylhexan-3-one in ref. 36, [α]D23 = - 30.8 (c 4, Et2O); for (R)-4-methylheptan-

3-one in ref. 37, [α]D25 = - 18 (c 1.5, hexane).

(39) G. Sudhakar, J. Raghavaiah, G. Mahesha and K. K. Singarapuc Org. Biomol. Chem. 2016,

14, 2866-2872.

(40) J. Liu and S. Ma Org. Lett. 2013, 15, 5150-5153.

(41) A. Z. Walton, PhD thesis, University of Florida, 2012.


24

(42) A. S. Abramovitz and V. Massey J. Biol. Chem. 1976, 251, 5321-5326.

(43) The Proteomics Handbook; eds. S. Duvaud, M. R. Wilkins, R. D. Appel and A. Bairoch,

Humana Press: Totowa, NJ, 2005.

(44) M. D. Winn, C. C. Ballard, K. D. Cowtan, E. J. Dodson, P. Emsley, P. R. Evans, R. M.

Keegan, E. B. Krissinel, A. G. W. Leslie, A. M. McCoy, S. J., G. N. Murshudov, N. S.

Pannu, E. A. Potterton, H. R. Powell, R. J. Read, A. Vagin and K. S. Wilson Acta

Crystallogr. D Biol. Crystallogr. 2011, 67, 235-242.

(45) P. D. Adams, P. V. Afonine, G. Bunkoczi, V. B. Chen, I. W. Davis, N. Echols, J. J. Headd,

L. W. Hung, G. J. Kapral, R. W. Grosse-Kunstleve, A. J. McCoy, N. W. Moriarty, R.

Oeffner, R. J. Read, D. C. Richardson, J. S. Richardson, T. C. Terwilliger and P. H. Zwart

Acta Crystallogr. D Biol. Crystallogr. 2010, 66, 213-221.

(46) P. Emsley, B. Lohkamp, W. G. Scott and K. Cowtan Acta Crystallogr. D Biol. Crystallogr.

2010, 66, 486-501.

(47) D. A. Case, R. M. Betz, W. Botello-Smith, D. S. Cerutti, T. E. Cheatham, III, T. A. Darden,

R. E. Duke, T. J. Giese, H. Gohlke, A. W. Goetz, N. Homeyer, S. Izadi, P. Janowski, J.

Kaus, A. Kovalenko, T. S. Lee, S. LeGrand, P. Li, C. Lin, T. Luchko, R. Luo, B. Madej, D.

Mermelstein, K. M. Merz, G. Monard, H. Nguyen, H. T. Nguyen, I. Omelyan, A. Onufriev,

D. R. Roe, A. Roitberg, C. Sagui, C. L. Simmerling, J. Swails, R. C. Walker, J. Wang, R.

M. Wolf, X. Wu, L. Xiao, D. M. York and P. A. Kollman, AMBER University of

California, San Francisco, 2016.

(48) C. Schneider and J. Sühnel Biopolymers 1999, 50, 287-302.

(49) W. L. Jorgensen, J. Chandraskhar, J. D. Madura, R. W. Impey and M. L. Klein J. Chem.

Phys. 1983, 79.


25

(50) S. Miyamoto and P. Kollman J. Comput. Chem. 1992, 13, 952-962.

(51) J.-P. Ryckaert, G. Ciccotti and H. J. C. Berendsen J. Comput. Phys. 1977, 23, 327-341.

(52) C. W. Hopkins, S. Le Grand, R. C. Walker and A. E. Roitberg J. Chem. Theory Comput.

2015, 11, 1864-1874.

(53) D. J. Sindhikara, S. Kim, A. F. Voter and A. E. Roitberg J. Chem. Theory Comput. 2009, 5,

1624-1631.

(54) S. Le Grand, A. W. Götz and R. C. Walker Comput. Phys. Commun. 2013, 184, 374-380.

(55) NADPD was prepared in situ by the NADP+-dependent oxidation of iso-propanol-d8 using

Thermoanaerobium brockii alcohol dehydrogenase.

(56) A. Z. Walton, W. C. Conerly, Y. Pompeu, B. Sullivan and J. D. Stewart ACS Catalysis

2011, 1, 989-993.

(57) B. Sullivan, A. Z. Walton and J. D. Stewart Enzyme Microb. Technol. 2013, 53, 70-77.

(58) E. Brenna, F. G. Gatti, A. Manfredi, D. Monti and F. Parmeggiani Catal. Sci. Technol.

2013, 3, 1136-1146.

(59) M. W. Fraaije and A. Mattevi Trends Biochem. Sci. 2000, 25, 126-132.

(60) A. B. Daughterty, J. R. Horton, X. D. Cheng and S. Lutz ACS Catalysis 2015, 5, 892-899.

(61) G. Tasnádi, C. K. Winkler, D. Clay, N. Sultana, W. M. F. Fabian, M. Hall, K. Ditrich and

K. Faber Chem. Eur. J. 2011, 18, 10362-10367.

(62) C. Stueckler, C. K. Winkler, M. Hall, B. Hauer, M. Bonnekessel, K. Zangger and K. Faber

Adv. Synth. Catal. 2011, 353, 1169-1173.


26

Figure Legends

Figure 1. Novel alkene substrates for OYE 3 and OYE 1. A. Trisubstituted alkenes 3 and 4

were designed as analogs of ethyl ketone 2 to test the sensitivity of OYE 3 and OYE to altered

steric demands in the β-position. B. 1H NMR spectrum of the racemic hydrogenated product

from enone 4. C. 2H NMR spectrum of the (4S,5R)-d2 product obtained by OYE 3-mediated

reduction of ketone 4 in D2O with NADPD (top) and the (4R,5S)-d2 product obtained by OYE 1-

mediated reduction of ketone 4 in D2O with NADPD (bottom). The peaks labelled by an asterisk

arise from acetone-d6 and iso-propanol-d8. .D. Steric demands of trisubstituted alkene substrates

of OYE 3 and OYE 1. The designations “L” and “S” refer to large and small substituents,

respectively.

Figure 2. Alkene substrates used to profile wild-type and Trp 116 site-saturation mutants

of OYE 3.

Figure 3. OYE 3 crystal structure. Electron density (2mFo – DFc) is shown at the 1.5 σ level.

A. Active site structure of crystals soaked with p-hydroxybenzaldehye. The positions of key

residues are shown. B. Loop 6 region (residues 289 – 309) of unsoaked OYE 3 crystals. Weak

electron density was observed for the side-chain of Glu 303. C. Loop 6 region of OYE 3

crystals soaked with p-hydroxybenzaldehyde. Weak electron density was observed beyond the

Cβ carbons for Glu 301, Glu 303 and Glu 306.


27

Figure 4. Structural alignment of loop 6 regions for unsoaked OYE 3 crystals (purple),

OYE 3 crystals soaked with p-hydroxybenzaldehyde (green) and OYE 1 crystals soaked

with p-hydroxybenzaldehyde (red). Distances from the Cα of residue 296 to N5 of FMN are ca.

10 Å in all cases.

Figure 5. MD simulation results for OYE 3 and OYE 1. A. Overlaid snapshots from a 100

ns simulation of OYE 3. Loop 6 is colored red and is the only portion of the protein structure

that undergoes significant movement. B. Overlaid snapshots from a 100 ns simulation of OYE 1

in the absence of substrate. Loop 6 is colored red. C. Distances between the Cα of residue 296

and N5 of FMN measured during MD simulations of OYE 3 (blue) and OYE 1 (red) in the

absence of substrate. D. Overlaid protein structures; only the loop 6 regions are shown: OYE 3

x-ray structure (green), OYE 3 average coordinates after MD simulation (blue), OYE 1 x-ray

structure (red), OYE 1 averaged coordinates after MD simulation (orange).


28

Table 1. Alkenes reduced by S. pastorianus OYE 1 and OYE 3 with differing

stereoselectivities. All substrates are shown in the “normal” orientation and the phenol inhibitor skeleton is indicated by dashed grey lines. Products derived from a “flipped” substrate binding orientation are shown in red.

Entry Substrate OYE 1 OYE 3

Ref. % ee % ee

1

75% (R) 59% (S) 61

2

60% (S)a 68% (R) 62

3

59% (R) 89% (S) 30

4

20% (R) 95% (S) 30

5

60% (R) 81% (S) 30

aThe indicated binding orientations presume that the activating ester group was the same for both OYE 1 and OYE 3; however, a change in activating ester group would also be consistent with the observed data.


29

Table 2. Reductions of trisubstituted olefins by OYE 3 and OYE 1. Reactions that yield products from “flipped” substrate binding orientations are shown in red, and key results are highlighted by grey shading. In entry the first number refers to fractional conversion and the second to % ee.

Amino acid at

position 116

%c a, % ee b

OYE 3 OYE 1 OYE 3 OYE 1

W (wt) >98%, 78% (S) >98%, 44% (R) 50%, 70% (S) >98%, 56% (R)

A 82%, >99% (S) 37%, 82% (S) ---c 5%, 58% (S)

C 32%, 96% (S) 40%, 70% (S) --- ---

D --- --- --- ---

E --- --- --- ---

F 32%, 14% (R) 16%, 65% (R) --- 58%, >98% (R)

G --- --- --- ---

H --- 9%, 40% (R) --- 8%, 80% (R)

I --- 59%, 0% --- 64%, 50% (R)

K --- --- --- ---

L >98%, 60% (S) >98%, 50% (R) 45%, 0% 14%, 78% (R)

M 76%, 36% (S) 57%, 67% (R) 47%, 0% 78%, >99% (R)

N 35%, 74% (S) 33%, 34% (R) 10%, 46% (S) 11%, 64% (R)

P --- 9%, 88% (R) --- ---

Q --- 20%, 82% (R) --- 18%, >99% (R)

R --- --- --- ---

S --- --- --- ---

T 82%, >99% (S) 9%, 67% (S) 6%, 99% (S) ---

V 20%, >99% (S) >98%, 86% (S) 31%, 96% (S) >98%, 54% (S)

Y 25%, 76% (S) 11%, 78% (R) 30%, 4% (R) >98%, 60% (R)

a Conversions were calculated on the basis of GC analysis of the crude mixture after 24 h reaction time (three replicates). b Optical purities were calculated from chiral-phase GC analysis (three replicates). c≤ 5% substrate conversion after 24 hr was observed.


30

Table 3. Trp 116 site-saturation mutagenesis data for OYE 3 and OYE 1. Reactions that yield products from “flipped” substrate binding orientations are shown in red, and key results are highlighted by grey shading.

Amino acid at

position 116

% ee %de

OYE 3 OYE 1a OYE 3 OYE1a OYE 3 OYE 1b OYE 3 OYE 1b

W (wt) ---c 60% (R) --- --- 76% (cis) 90% (cis) 75% (trans) 97% (trans)

A >98% (S) 72% (S) >98% (S) >98% (S) --- 93% (trans) 8% (cis) 56% (cis)

C --- 77% (S) >98% (S) >98% (S) 67% (cis) 73% (trans) --- ---

D --- 77% (S) --- 91% (S) 76% (cis) >98% (cis) 75% (trans) 94% (trans)

E 23% (S) 88% (S) 2% (R) 90% (S) 84% (cis) 86% (trans) 78% (trans) ---

F 59% (S) >98% (S) >98% (S) >98% (S) 81% (cis) 46% (cis) 66% (trans) 97% (trans)

G >98% (S) 86% (S) >98% (S) >98% (S) 81% (cis) 93% (trans) 75% (trans) ---

H 71% (S) 77% (S) 78% (S) >98% (S) 81% (cis) 75% (cis) 71% (trans) 96% (trans)

I --- 91% (S) --- >98% (S) 83% (cis) 91% (trans) 88% (trans) 73% (trans)

K --- 76% (S) --- >98% (S) 75% (cis) --- 73% (trans) 95% (trans)

L 70% (S) 57% (S) >98% (S) >98% (S) 67% (cis) 60% (cis) 60% (trans) 96% (trans)

M >98% (S) 86% (S) >98% (S) >98% (S) 74% (cis) 78% (trans) 44% (trans) 89% (trans)

N 62% (S) 89% (S) >98% (S) >98% (S) 77% (cis) 70% (trans) 60% (trans) 69% (trans)

P --- 77% (S) --- >98% (S) 79% (cis) 38% (cis) 79% (trans) 86% (trans)

Q 71% (S) 89% (S) >98% (S) >98% (S) 70% (cis) 82% (trans) 52% (trans) 47% (trans)

R --- --- --- --- >98% (cis) --- 70% (trans) ---

S --- >98% (S) >98% (S) >98% (S) 71% (cis) >98% (trans) 68% (trans) ---

T >98% (S) >98% (S) >98% (S) >98% (S) 64% (cis) 85% (trans) 25% (trans) ---

V 53% (S) 92% (S) >98% (S) >98% (S) 43% (cis) >98% (trans) 6% (trans) 56% (cis)

Y 73% (S) 87% (S) >98% (S) >98% (S) 81% (cis) >98% (cis) 71% (trans) >98% (trans)

aRef. 56 bRef. 31 c≤ 10% substrate conversion after 24 hr was observed.


31

Table 4. Summary of x-ray crystallography data.

Structure title Wild-type OYE 3 p-HBA-soaked OYE 3

PDB ID 5V4V 5V4P

Ligand Soaked none p-hydroxybenzaldehyde

X-Ray Source 21-ID-G beamline 21-ID-G beamline

APS Argonne National

Laboratory APS Argonne National

Laboratory

Space Group P 21 21 21 P 21 21 21

Unit Cell Dimensions

a, b, c (Å) 61.2 107.8 141.1 61.5 106.4 141.0

Resolution (Å) 38.9 - 1.80

(1.86 - 1.80)a

24.9 - 1.88

(1.95 - 1.88)

Unique Reflections 87273 (8428) 76003 (7463)

Completeness% 99.8 (97.8) 100 (99.8)

Multiplicity 14.5 (14.3) 14.6 (14.7)

CC1/2 0.999 (0.964) 0.999 (0.978)

I/σ (I) 20.2 (3.95) 18.6 (5.36)

Rwork 0.148 (0.230) 0.181 (0.205)

Rfree 0.189 (0.269) 0.219 (0.251)

Ramachandran Statistics

Favored (%) 97 97

Allowed (%) 2.7 3.2

Outliers (%) 0 0

Number of Protein, Solvent, and Ligand

Atoms 6396, 930, 92 6325, 396, 126

Average B Factors (Å2)

21.6 18.5

Protein 20.1 18.2

Ligands 22.4 25.1

Solvent 32.5 22.3

aValues in parentheses denote data for the highest resolution bin.


32

A

Figure 1

CH3 CH3

O

CH3

H4

H5

H5

B


33

CH3 CH3

O

CH3

D

HD

45OYE 3

(flipped)

OYE 1

(normal)

(4S,5R)-d2

(4R,5S)-d2

C

Figure 1


34

D

Figure 1


35

Figure 2


36

Figure 3

Asn 194 His

191

FMN

Trp 116

Leu 118

Ser 296

p-hydroxy-

benzaldehyde

A


37

Figure 3

Ser

296

Glu

301 Glu

306

*

* Glu 303

B


38

Figure 3

Ser

296

Glu

301

Glu

306

*

* Glu 303

C


39

Ser

296

Phe

296

Figure 4


40

A B

Figure 5


41

Figure 5

C


42

Figure 5

D


43


investigating saccharomyces cerevisiae alkene reductase

Documents